Purpose: Few studies have investigated the prognostic significance of perineural invasion (PNI) in gastric cancer. Therefore, we examined the association between PNI and clinicopathological factors and the effect of PNI on overall survival in patients with gastric carcinoma.
Methods: Paraffin sections of surgical specimens from 354 patients who underwent gastric resection were stained with hematoxylin and eosin. PNI was assessed histologically as positive when cancer cells were seen in the perineurium or neural fascicles. Survival analysis was done in 219 patients with T(2,3,4) tumors who underwent potentially curative resection. Data were collected prospectively.
Results: PNI was positive in 211 of the 354 patients (59.6%). The ratio of undifferentiated tumors, tumors with vascular invasion, and lymph node metastasis was significantly higher in the PNI-positive patients than in the PNI-negative patients ( P < 0.0001 for all three associations). As the depth of mural invasion increased, so did PNI positivity ( P < 0.0001). The overall survival of the PNI-positive patients was significantly worse than that of the PNI-negative patients in the univariate analysis ( P = 0.0009). However, PNI had no independent prognostic significance in the multivariate Cox proportional hazards model analysis. When the patients were separated into subgroups, PNI had prognostic value in patients with T(3) tumors ( P = 0.036) and no lymph node metastasis ( P = 0.005) in the univariate analysis, but no prognostic significance in the multivariate analysis.
Conclusions: Although the incidence of PNI is high in gastric carcinoma and increases with the progression of disease, it does not provide any additional information to the classical prognostic parameters.